---
dataset_info:
  features:
  - name: en_text
    dtype: string
  - name: az_text
    dtype: string
  splits:
  - name: train
    num_bytes: 327668699
    num_examples: 1391227
  download_size: 204322794
  dataset_size: 327668699
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
task_categories:
- translation
- sentence-similarity
language:
- en
- az
tags:
- azerbaijani
- english
- parallel
- corpus
pretty_name: English Azerbaijani Parallel Corpus
size_categories:
- 1M<n<10M
license: cc-by-4.0
---

# English–Azerbaijani Parallel Corpus

This dataset contains ~1,391,230 sentence pairs translated from English into Azerbaijani.

## Sources (English side)

The English sentences were sampled and cleaned from six native-English corpora across complementary registers — written prose, encyclopedic and news text, plus conversational and instructional registers:

| Source | Register | Pairs | Share | Link |
| --- | --- | --- | --- | --- |
| FineWeb-Edu (100B shuffle) | general web prose | 400,000 | 28.75% | https://huggingface.co/datasets/karpathy/fineweb-edu-100b-shuffle |
| English Wikipedia | encyclopedic | 400,000 | 28.75% | https://huggingface.co/datasets/raj2708/wikipedia-en |
| UltraChat 200k | instructional / chat prompts | 250,547 | 18.01% | https://huggingface.co/datasets/HuggingFaceH4/ultrachat_200k |
| News-Commentary (en side) | news / economics | 148,600 | 10.68% | https://huggingface.co/datasets/Helsinki-NLP/news_commentary |
| DailyDialog | everyday dialogue | 108,118 | 7.77% | https://huggingface.co/datasets/roskoN/dailydialog |
| Subtitles | spoken / subtitle dialogue | 83,965 | 6.04% | https://opus.nlpl.eu/datasets/OpenSubtitles?pair=az&en) |

Total: 1,391,230 pairs.

## Columns

- `en_text` — English source sentence
- `az_text` — Azerbaijani translation

## License

Sentences originate from multiple corpora, each under its own license. Downstream use
must respect the license of every underlying source. (NOTE: the previous CC-BY-4.0 tag
no longer holds across all sources — verify each, especially the subtitle source.)

## Contact

For more information, questions, or issues, please contact LocalDoc at [v.resad.89@gmail.com].